Detecting the extent of control over selection bias relating to oral health and otorhinolaryngology: cross-sectional study